Limitations of Epidemiological and Serologic Studies Addressing Simian Virus 40 and Non-Hodgkin's Lymphoma

Sir: I read with interest the article titled “Does Simian Virus 40 Cause Non-Hodgkin's Lymphoma? A Review of the Laboratory and Epidemiological Evidence” ([1]), in which the author reviewed the epidemiological and serologic data of polyomavirus simian virus 40 (SV40) and its link to non-Hodgkin's lymphoma (NHL). However, before causal conclusions are made ([1]), it is important to point out that several limitations have been recognized for the epidemiological ([2],[3],[4]) and serologic ([5], [6]) studies addressing SV40 and human malignancies. Indeed, an evaluation by the Institute of Medicine Immunization and Safety Review Committee found that the design and data of epidemiological studies to examine cancer rates in people exposed to SV40-contaminated vaccines are inadequate to evaluate a causal relationship ([2]). This is significant because the epidemiological studies examining SV40 and NHL ([7],[8],[9],[10],[11]) have the same design and analysis as those evaluated by the Institute of Medicine ([2]). In addition, the validity of observational studies, such as the retrospective epidemiological and cross-sectional serologic studies cited by Engels ([1]), depends on the accuracy of the existing knowledge of the biological properties of SV40 and the identification of the human population infected with the virus ([2],[3],[4]). Therefore, supportive evidence from experimental studies is required in order to draw causal inferences in human disease ([4], [12]).

A known confounding factor in the epidemiological studies of SV40 and NHL ([1]) is that the actual number of people infected with live SV40 through the use of contaminated polio vaccines in the United States or other countries is not known ([2],[3],[4]). It is recognized that not all vaccine lots were contaminated with SV40, formalin inactivation was expected to reduce the titer of live SV40 in the lots that were contaminated, and successful infection rates by live SV40 are unknown ([2],[3],[4]). For example, only 19 percent of newborn children and 15 percent of infants 3–6 months old at the time of receiving a known contaminated oral poliovaccine were shown to excrete infectious SV40 in their stools for up to 5 weeks after vaccination ([13]), indicating established infection. Therefore, an inability to identify the population actually infected with SV40 through the use of contaminated polio vaccines precludes a meaningful calculation of cancer incidence in relation to exposure to those vaccines. More importantly, the detection and occasional isolation of SV40 from children and adult patients who could not have been infected with SV40 through contaminated vaccines suggest that SV40 infections are continuing to occur at the present time in different populations and are not limited to the recipients of those contaminated vaccines ([2], [4], [5]). Further complication for epidemiological analysis is that recently it has been shown that some SV40-contaminated oral polio vaccines were produced and may have been used throughout the world until 1978 ([14]).

Studies of malignant neoplasias, including lymphomas by SV40 in hamsters, have indicated that the carcinogenesis process is multifactorial. The outcome is influence by host-dependent factors, such as age (newborn vs. weanling vs. adult), route of inoculation (intracerebral vs. subcutaneous vs. intravenous), and target cell type (brain, mesothelia, lymphoid, etc.) ([5]). A virus-dependent factor, dose of inoculum, is also a predictor of the development of cancer ([5]). Recent studies indicated that strain distinction may represent another factor that influences the dynamics of infection by SV40 ([15]), including the development of malignancies. The results also showed that not all hosts exposed to large doses of SV40 developed detectable levels of long-lasting antibody (Table 1).

Table 1 Tumor development and antibody responses to SV40 in Syrian golden hamsters inoculated as weanlings with strains VA45-54 and SVCPC^a

		No. tumors/No. hamsters injected (%)	Neutralizing antibody^b
SV40 strain	Tumors	Injected (%)	No. positive/No, tested (%)
VA45-54^d	Yes	7/33 ([21])	7/7 (100)
	No	26/33 (79)	10/26 ([38])
	Totals		17/33 (52)
SVCPC^e	Yes	18/37 (49)	18/18 (100)
	No	19/37 (51)	12/13 (92)
	Totals		30/31 (97)
Controls	No	0/58 (0)	0^c(0)

^aSummarized from Vilchez et al. ([15]).

^bSera were collected when tumor-bearing animals were sacrificed or when the experiments were terminated. A plaque-reduction assay was used to measure neutralizing antibodies against SV40.

^cn = 19 samples tested.

^dVirus strain isolated from monkeys.

^eVirus strain isolated and detected in human malignancies including lymphomas.

Serological studies in humans have not yet provided certain SV40 seroprevalence data, but seropositivity rates appear to be low. Retrospective cross-sectional studies of stored serum samples using ELISA tests based on virus-like particles have encountered cross-reactive JCV and BKV antibodies at low serum dilutions that have hampered estimates of SV40 infections ([16],[17],[18],[19],[20],[21],[22]), leading to some conclusions that there was no evidence of human infections by SV40 or of an association of SV40 with lymphomas ([16], [19],[20],[21],[22]). Polyomavirus serological surveys are complicated. ELISA tests can detect nonneutralizing antibodies and some cross-reactivity is not unexpected as the 3 viruses (BKV, JCV, SV40) are partially related ([23]). This is important as those assays were directed against the polyomavirus VP1 protein, a highly homologous gene product among the three viruses. Human immune responses to SV40 infections have not been characterized; specific antibody responses may be low-titered and, as indicated recently, may wane over time ([24]). It is worth noting that children excreting infectious SV40 in their stools after receiving a known contaminated oral poliovaccine failed to show neutralizing antibody responses to the viral infection ([13]) and a person with a proven SV40-positive brain tumor did not possess detectable SV40 antibodies ([25]).

Modern molecular biology assays known to be highly sensitive and specific are excellent alternatives for the analysis of SV40 infections and NHL. Indeed, SV40 markers in lymphomas from adult patients have been reported in several independent studies in the United States ([26],[27],[28],[29],[30],[31],[32],[33]), Italy ([34]), Japan ([35]), and Costa Rica ([36]) as well as in pediatric lymphoid cancers of German origin ([37]). Furthermore, the expression of the viral oncoprotein ([32],[33],[34], [36]), the detection of SV40 DNA in many malignant B-cells by in situ hybridization ([35]) and the specific methylation of several promoter genes in SV40-positive NHL ([31]) supports the possibility that the virus may be functionally important in some of those human lymphomas. However, a report from Japan ([38]) found a very low frequency of SV40-positive NHL and studies from Western Europe ([39],[40],[41],[42]) failed to detect SV40 sequences in systemic lymphomas. These differences among studies may occur because of unrecognized differences among specimens available for analysis, technical variations in sample processing or assay ([5], [43]), or because of the inclusion of many T-cell neoplasias and Epstein Barr virus (EBV)-positive lymphomas in surveys ([38],[39],[40]). The latter circumstances may reduce the probability of detecting SV40-positive tumors because SV40 rarely has been found in T-cell cancers and SV40 and EBV seldom are found together in a given tumor ([28], [29], [31]). However, more important than technical issues in affecting differences between studies is that the prevalence of SV40 infections may vary among different geographic regions ([2], [5]). This may reflect the history of usage of contaminated poliovaccines within a given area ([2], [4], [5], [14]). Unfortunately, records are frequently unavailable or inadequate to identify the frequency or duration of use of contaminated vaccines within a certain region, or the population groups most exposed to those contaminated vaccines ([2], [4], [5]). However, as lymphomas can arise from different causal factors, variation in the frequency of detection of SV40 in human lymphoproliferative disorders among population groups does not rule out a possible functional role for the virus in those malignancies in which it is present.

Current molecular biology, pathology, and clinical data, taken together, show that SV40 is significantly associated with and may be functionally important in the development of some human lymphomas. In addition, data support the conclusion of the Institute of Medicine of the National Academies that SV40 exposure could lead to cancer in humans under natural conditions ([2]). Future studies using informative and specific methodology need to focus on how SV40 is transmitted among humans, how it is distributed throughout the infected host, and how the host responds immunologically.

Conflict of Interest Statement: None declared.